How to Write Better AI Prompts
(Complete Beginner's Guide)

If you've used ChatGPT and felt underwhelmed by the results, you're not alone — and it's almost certainly not ChatGPT's fault. The most common reason people get mediocre AI outputs is simple: they're asking vague questions and getting vague answers.

Think about it this way: if you walked up to a world-class expert and said "tell me about marketing," you'd get a generic overview. But if you said "I run a Shopify store selling handmade dog collars with $500/month to spend on ads, targeting millennial dog parents in the US — what's the most cost-effective way to acquire my first 100 customers?" you'd get something actually useful.

That's the fundamental principle of prompt writing: the quality of your output is directly proportional to the quality of your input. This guide will teach you exactly how to make your inputs much, much better.

What Is a Prompt, Exactly?

A prompt is simply the instruction or question you give to an AI. It can be as short as "write a poem" or as detailed as a multi-paragraph brief with role assignments, constraints, format requirements, and examples. Both are valid — but they produce very different results.

Good prompts share a few key characteristics: they're specific, they provide context, they define the desired output format, and they often assign the AI a role to play. We'll break down each of these elements in detail throughout this guide.

The CRATE Framework: A Simple System for Better Prompts

After testing hundreds of prompt formats, the best mental model for writing strong prompts is what I call the CRATE framework. It stands for: Context, Role, Action, Tone, and Examples.

C — Context

Context is the background information the AI needs to understand your situation. Without context, the AI has to make assumptions — and those assumptions are often wrong. Good context answers: Who are you? What are you working on? Who is the audience? What constraints or limitations exist?

Example without context: "Write a social media post about my product."
Example with context: "I run a small handmade candle brand on Etsy targeting women aged 25–35 who value sustainable, natural products. Write an Instagram caption for a new lavender soy wax candle priced at $28."

The second prompt gives the AI your business type, platform, audience demographics, values, product specifics, and price — all of which shape the output dramatically.

R — Role

Assigning the AI a role is one of the single most effective prompt techniques there is. When you say "You are a [specific expert]," you're telling the AI to draw on a specific subset of its knowledge and apply a particular expertise lens to your request.

The difference in output quality is remarkable. Role prompts work especially well for: writing tasks (copywriter, journalist, author), analysis tasks (consultant, analyst, strategist), and learning tasks (teacher, tutor, coach).

A — Action

The action is the specific thing you want the AI to do. Be precise about the verb. "Write," "analyze," "compare," "list," "summarize," "rewrite," "create" — each produces a different kind of output. Vague verbs like "help me with" or "tell me about" produce vague results.

Also specify the scope and length. "Write a blog post" could mean 300 words or 3,000 words. "Write a 900-word blog post structured as an introduction, 4 sections with H2 headings, and a conclusion" gives the AI everything it needs.

T — Tone

Tone is critically underused in most prompts. The same information written in a "formal and authoritative" tone vs. a "conversational and friendly" tone reads completely differently — and one might be exactly right for your audience while the other falls flat.

Good tone descriptors include: formal, conversational, inspirational, urgent, empathetic, humorous, direct, storytelling, educational, bold, minimalist. Combine 2-3 for nuance: "professional but approachable" or "energetic and slightly irreverent."

E — Examples

When you provide examples of what you want (or don't want), AI performance jumps significantly. This technique is called "few-shot prompting" in the research world, but in practice it just means showing the AI one or two examples of the output you're looking for.

The 5 Most Common Prompt Mistakes (and How to Fix Them)

Mistake 1: Asking for Everything at Once

Beginners often try to cram 10 different requests into one prompt. The AI becomes overwhelmed and produces shallow output for each. Instead, break complex tasks into a series of focused prompts. First get the structure, then ask it to flesh out each section.

Mistake 2: Not Specifying Format

By default, ChatGPT often produces long paragraphs. If you want bullet points, a numbered list, a table, a JSON structure, or a specific document format — say so explicitly. "Format your response as a bulleted list" or "Present this as a table with three columns: Pros, Cons, Verdict."

Mistake 3: Accepting the First Draft

Great prompting is iterative. The first output is a starting point, not a final product. Follow up with refinements: "Make the tone more casual," "Expand the third section with two more examples," "Rewrite the headline to focus on the emotional benefit rather than the feature."

Mistake 4: Ignoring the "Don'ts"

Telling the AI what NOT to do is just as important as telling it what to do. "Don't use jargon," "Don't start with 'I'," "Don't include generic filler phrases like 'in today's fast-paced world'" — these constraints prevent the most common AI writing tics.

Mistake 5: Being Too Polite

You don't need to say "please" or "thank you" to an AI (it doesn't have feelings), and excessive politeness often dilutes your prompt. Be clear and direct: "List 10 product names" not "Could you possibly help me think about some product name ideas?"

Advanced Prompt Techniques for Better Results

Chain-of-Thought Prompting

For complex analytical tasks, adding "Think step by step" or "Before answering, explain your reasoning process" dramatically improves accuracy. This works because it forces the AI to work through logic systematically rather than jumping to a conclusion.

Persona Layering

Go beyond basic roles by layering specific expertise: "You are a copywriter who worked at Ogilvy for 10 years, specializing in direct response ads for financial products targeting retirees." The more specific the persona, the more tailored the output.

Negative Prompting

Tell the AI what to avoid: "Do not include generic advice that could apply to any business. Every suggestion must be specific to [MY INDUSTRY]. Do not recommend strategies that require a budget over $500."

Output Format Templates

Instead of describing the format in words, show it with a template:

Templates like this produce consistently formatted outputs that are easy to scan and immediately usable — no extra editing required at all.

Building Your Personal Prompt Library

One of the most valuable things you can do as an AI user is build and maintain your own collection of high-performing prompts. Every time you get an output you love, save that prompt. Over time, you'll develop a personal library of tested, proven prompts tailored to your specific use cases. You will never start from scratch again.

At MyPromptsGenerator, we've done part of this work for you. Our free prompt library contains 500+ expert-crafted prompts across 8 categories that you can use as starting points and customize to your needs.

Putting It All Together: Before and After

Let's look at a real-world transformation from a weak prompt to a strong one using the CRATE framework:

Weak prompt: "Write me an email to my subscribers about my new course."

Strong prompt using CRATE:

The second prompt will produce an output that's dramatically better, more targeted, and closer to finished quality — because you gave the AI everything it needs to succeed.

Final Thoughts

Writing better prompts is a skill — and like any skill, it gets easier with practice. The CRATE framework gives you a solid foundation: start with context, assign a role, define the action precisely, specify the tone, and provide examples where possible.

Don't get discouraged if your first few attempts don't produce perfect results. The real skill in prompting is knowing how to iterate. Each output teaches you something about how to refine your prompt for next time.

As you get more comfortable, you'll develop an intuition for how AI thinks and what kinds of prompts unlock its best capabilities. Until then, start with proven templates from our library, adapt them to your needs, and build from there.

Ready to practice? Try our YouTube Script Prompt Generator to see these principles in action, or browse the full prompt library for copy-ready templates.

Related articles: ChatGPT YouTube Prompts · Dropshipping AI Prompts